SchemaCrawler Serialization

SchemaCrawler allows it’s underlying schema metadata model to be serialized into various formats that can be programatically consumed by other applications and systems. SchemaCrawler’s native serialization format uses Java serialization, and is tied to a given version of Java and a given version of SchemaCrawler. This type of serialization is most useful when using the SchemaCrawler Interactive Shell, since you can save off the schema metadata model to be used in a later shell session, in which case you do not need an active connection to the database. SchemaCrawler does not offer deserialization using formats other than Java serialization.

SchemaCrawler needs Jackson jars on the classpath to serialize to JSON and YAML, and these can be downloaded with the download tool provided with the SchemaCrawler download.

How to Serialize a Catalog

SchemaCrawler serialization can be run using the --command=serialize command-line option. The serialized schema metadata model will be saved to a file.

For more details, see the serialize example in the SchemaCrawler examples download.

SchemaCrawler serialization can produce output in binary Java serialization, JavaScript object notation (JSON) or YAML formats. (Click on the links for example output.) A serialized schema metadata model will be produced in the format specified using the --output-format command-line option. For example, --output-format=json will generate a output in JSON format.

JSON and YAML formats cannot be diff-ed since they contain GUIDs, and the formats are specific to a given version of SchemaCrawler, and subject to change from version to version.

How to Load a Serialized Catalog

When you use Java serialization (--output-format=ser) to serialize database metadata to a file, you can load a serialized metadata using the offline server type. Please make sure that you use the same version of SchemaCrawler to create serialized snapshots and to load those offline catalog snapshots.